AMD Corporate VP Madhu Rangarajan said there is no one-size-fits-all for agentic AI infrastructure, arguing that workloads labeled as such vary widely in memory, storage, and CPU needs. He noted that …
Drew Henry, EVP of Arm's Physical AI Business Unit, argues that robotics, humanoids, and autonomous vehicles represent one of the hardest systems-engineering problems in computing, with the physical e…
AMD agreed to acquire Taalas, a 25-person Toronto-based chip startup that builds model-specific silicon, for an undisclosed amount. Taalas's first chip, the HC1, runs only Llama 3.1 8B, with weights h…
TSMC reported a fifth straight record gross margin of 67.7% in Q2 2026, but guided Q3 down to 66.0%, while gross profit rose to roughly $27.2 billion. Despite calling AI demand unbounded, TSMC is spen…
PicoJool CEO Al Yuen argues that GaAs VCSELs at 200G/lane are the pragmatic choice for scaling optical interconnects, citing unconstrained supply versus indium phosphide and a roadmap to 12.8T via 8×2…
Two AI chip startups, Groq and Cerebras, achieved massive financial outcomes in the past year—Groq was acquired for $20 billion and Cerebras IPO'd to a $40 billion-plus market cap—but their pre-LLM ar…
High Bandwidth Flash (HBF), a NAND-based memory stack with read bandwidth comparable to HBM4 but 10x the capacity, is set to sample from Sandisk in late 2026 and appear in AI inference devices by earl…
AMD, Intel, Nvidia, Arm, and Qualcomm are selling datacenter CPUs into the AI buildout, but the instruction set architecture (ISA) matters less than the coherent link to the GPU. The coherent host soc…
AMD, Intel, Nvidia, Arm, and Qualcomm are competing in the datacenter CPU market driven by agentic AI, which has shifted the CPU-to-GPU ratio from 1:4 toward 1:1. The market sees these CPUs as commodi…
Nvidia's Deepu Talla discusses the state of physical AI and robotics, highlighting that the industry is still in early stages with only a million or two robots shipped annually against a tens-of-billi…